github/DeepFaceLab
1
0
Fork 0
mirror of https://github.com/iperov/DeepFaceLab.git synced 2025-08-22 06:23:20 -07:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge pull request #12 from MachineEditor/config_files_updates

Browse source 
Config files updates
This commit is contained in:
Ognjen 2021-12-09 19:12:57 +01:00 • committed by GitHub
commit 98e1f7c31f
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
6 changed files with 113 additions and 55 deletions
Download patch file Download diff file Expand all files Collapse all files

77
models/Model_AMP/Model.py
View file

@ -46,7 +46,6 @@ class AMPModel(ModelBase):
default_lr_dropout = self.options['lr_dropout'] = self.load_or_def_option('lr_dropout', 'n') default_lr_dropout = self.options['lr_dropout'] = self.load_or_def_option('lr_dropout', 'n')
default_random_warp = self.options['random_warp'] = self.load_or_def_option('random_warp', True) default_random_warp = self.options['random_warp'] = self.load_or_def_option('random_warp', True)
default_random_hsv_power = self.options['random_hsv_power'] = self.load_or_def_option('random_hsv_power', 0.0)
default_random_downsample = self.options['random_downsample'] = self.load_or_def_option('random_downsample', False) default_random_downsample = self.options['random_downsample'] = self.load_or_def_option('random_downsample', False)
default_random_noise = self.options['random_noise'] = self.load_or_def_option('random_noise', False) default_random_noise = self.options['random_noise'] = self.load_or_def_option('random_noise', False)
default_random_blur = self.options['random_blur'] = self.load_or_def_option('random_blur', False) default_random_blur = self.options['random_blur'] = self.load_or_def_option('random_blur', False)
@ -58,6 +57,9 @@ class AMPModel(ModelBase):
default_random_color = self.options['random_color'] = self.load_or_def_option('random_color', False) default_random_color = self.options['random_color'] = self.load_or_def_option('random_color', False)
default_clipgrad = self.options['clipgrad'] = self.load_or_def_option('clipgrad', False) default_clipgrad = self.options['clipgrad'] = self.load_or_def_option('clipgrad', False)
default_usefp16 = self.options['use_fp16'] = self.load_or_def_option('use_fp16', False) default_usefp16 = self.options['use_fp16'] = self.load_or_def_option('use_fp16', False)
default_cpu_cap = self.options['cpu_cap'] = self.load_or_def_option('default_cpu_cap', 8)
default_preview_samples = self.options['preview_samples'] = self.load_or_def_option('preview_samples', 4)
default_lr_modifier = self.options['lr_modifier'] = self.load_or_def_option('lr_modifier', 0)
ask_override = False if self.read_from_conf else self.ask_override() ask_override = False if self.read_from_conf else self.ask_override()
if self.is_first_run() or ask_override: if self.is_first_run() or ask_override:
@ -66,12 +68,14 @@ class AMPModel(ModelBase):
self.ask_session_name() self.ask_session_name()
self.ask_maximum_n_backups() self.ask_maximum_n_backups()
self.ask_write_preview_history() self.ask_write_preview_history()
self.options['preview_samples'] = np.clip ( io.input_int ("Number of samples to preview", default_preview_samples, add_info="1 - 16", help_message="Typical fine value is 4"), 1, 16 )
self.ask_target_iter() self.ask_target_iter()
self.ask_retraining_samples() self.ask_retraining_samples()
self.ask_random_src_flip() self.ask_random_src_flip()
self.ask_random_dst_flip() self.ask_random_dst_flip()
self.ask_batch_size(8) self.ask_batch_size(8)
self.options['use_fp16'] = io.input_bool ("Use fp16", default_usefp16, help_message='Increases training/inference speed, reduces model size. Model may crash. Enable it after 1-5k iters.') self.options['use_fp16'] = io.input_bool ("Use fp16", default_usefp16, help_message='Increases training/inference speed, reduces model size. Model may crash. Enable it after 1-5k iters.')
self.options['cpu_cap'] = np.clip ( io.input_int ("Max cpu cores to use.", default_cpu_cap, add_info="1 - 256", help_message="Typical fine value is 0.5"), 1, 256 )
@ -106,6 +110,8 @@ class AMPModel(ModelBase):
if self.is_first_run() or ask_override: if self.is_first_run() or ask_override:
if (self.read_from_conf and not self.config_file_exists) or not self.read_from_conf: if (self.read_from_conf and not self.config_file_exists) or not self.read_from_conf:
morph_factor = np.clip ( io.input_number ("Morph factor.", default_morph_factor, add_info="0.1 .. 0.5", help_message="Typical fine value is 0.5"), 0.1, 0.5 ) morph_factor = np.clip ( io.input_number ("Morph factor.", default_morph_factor, add_info="0.1 .. 0.5", help_message="Typical fine value is 0.5"), 0.1, 0.5 )
self.options['morph_factor'] = morph_factor self.options['morph_factor'] = morph_factor
@ -121,6 +127,8 @@ class AMPModel(ModelBase):
self.options['blur_out_mask'] = io.input_bool ("Blur out mask", default_blur_out_mask, help_message='Blurs nearby area outside of applied face mask of training samples. The result is the background near the face is smoothed and less noticeable on swapped face. The exact xseg mask in src and dst faceset is required.') self.options['blur_out_mask'] = io.input_bool ("Blur out mask", default_blur_out_mask, help_message='Blurs nearby area outside of applied face mask of training samples. The result is the background near the face is smoothed and less noticeable on swapped face. The exact xseg mask in src and dst faceset is required.')
self.options['loss_function'] = io.input_str(f"Loss function", default_loss_function, ['SSIM', 'MS-SSIM', 'MS-SSIM+L1'], help_message="Change loss function used for image quality assessment.") self.options['loss_function'] = io.input_str(f"Loss function", default_loss_function, ['SSIM', 'MS-SSIM', 'MS-SSIM+L1'], help_message="Change loss function used for image quality assessment.")
self.options['lr_modifier'] = np.clip (io.input_int("Learningrate factor", default_lr_modifier, add_info="-100 .. 100", help_message="Modify the Learning rate: 100 == multipy by 4, -100 == divide by 4"), -100, 100)
self.options['lr_dropout'] = io.input_str (f"Use learning rate dropout", default_lr_dropout, ['n','y','cpu'], help_message="When the face is trained enough, you can enable this option to get extra sharpness and reduce subpixel shake for less amount of iterations. Enabled it before `disable random warp` and before GAN. \nn - disabled.\ny - enabled\ncpu - enabled on CPU. This allows not to use extra VRAM, sacrificing 20% time of iteration.") self.options['lr_dropout'] = io.input_str (f"Use learning rate dropout", default_lr_dropout, ['n','y','cpu'], help_message="When the face is trained enough, you can enable this option to get extra sharpness and reduce subpixel shake for less amount of iterations. Enabled it before `disable random warp` and before GAN. \nn - disabled.\ny - enabled\ncpu - enabled on CPU. This allows not to use extra VRAM, sacrificing 20% time of iteration.")
default_gan_power = self.options['gan_power'] = self.load_or_def_option('gan_power', 0.0) default_gan_power = self.options['gan_power'] = self.load_or_def_option('gan_power', 0.0)
@ -142,8 +150,6 @@ class AMPModel(ModelBase):
self.options['random_blur'] = io.input_bool("Enable random blur of samples", default_random_blur, help_message="") self.options['random_blur'] = io.input_bool("Enable random blur of samples", default_random_blur, help_message="")
self.options['random_jpeg'] = io.input_bool("Enable random jpeg compression of samples", default_random_jpeg, help_message="") self.options['random_jpeg'] = io.input_bool("Enable random jpeg compression of samples", default_random_jpeg, help_message="")
self.options['random_hsv_power'] = np.clip ( io.input_number ("Random hue/saturation/light intensity", default_random_hsv_power, add_info="0.0 .. 0.3", help_message="Random hue/saturation/light intensity applied to the src face set only at the input of the neural network. Stabilizes color perturbations during face swapping. Reduces the quality of the color transfer by selecting the closest one in the src faceset. Thus the src faceset must be diverse enough. Typical fine value is 0.05"), 0.0, 0.3 )
self.options['gan_power'] = np.clip ( io.input_number ("GAN power", default_gan_power, add_info="0.0 .. 5.0", help_message="Forces the neural network to learn small details of the face. Enable it only when the face is trained enough with random_warp(off), and don't disable. The higher the value, the higher the chances of artifacts. Typical fine value is 0.1"), 0.0, 5.0 ) self.options['gan_power'] = np.clip ( io.input_number ("GAN power", default_gan_power, add_info="0.0 .. 5.0", help_message="Forces the neural network to learn small details of the face. Enable it only when the face is trained enough with random_warp(off), and don't disable. The higher the value, the higher the chances of artifacts. Typical fine value is 0.1"), 0.0, 5.0 )
@ -196,7 +202,6 @@ class AMPModel(ModelBase):
morph_factor = self.options['morph_factor'] morph_factor = self.options['morph_factor']
gan_power = self.gan_power = self.options['gan_power'] gan_power = self.gan_power = self.options['gan_power']
random_warp = self.options['random_warp'] random_warp = self.options['random_warp']
random_hsv_power = self.options['random_hsv_power']
if 'eyes_mouth_prio' in self.options: if 'eyes_mouth_prio' in self.options:
self.options.pop('eyes_mouth_prio') self.options.pop('eyes_mouth_prio')
@ -384,6 +389,13 @@ class AMPModel(ModelBase):
self.model_filename_list += [ [self.GAN, 'GAN.npy'] ] self.model_filename_list += [ [self.GAN, 'GAN.npy'] ]
# Initialize optimizers # Initialize optimizers
lr_modifier = self.options['lr_modifier']
if lr_modifier == 0:
lr = 5e-5
elif lr_modifier < 0:
lr = 5e-5 / abs( lr_modifier * 4/100 )
else:
lr = 5e-5 * abs( lr_modifier * 4/100 )
clipnorm = 1.0 if self.options['clipgrad'] else 0.0 clipnorm = 1.0 if self.options['clipgrad'] else 0.0
if self.options['lr_dropout'] in ['y','cpu']: if self.options['lr_dropout'] in ['y','cpu']:
lr_cos = 500 lr_cos = 500
@ -394,17 +406,17 @@ class AMPModel(ModelBase):
self.G_weights = self.encoder.get_weights() + self.decoder.get_weights() self.G_weights = self.encoder.get_weights() + self.decoder.get_weights()
OptimizerClass = nn.AdaBelief if adabelief else nn.RMSprop OptimizerClass = nn.AdaBelief if adabelief else nn.RMSprop
self.src_dst_opt = OptimizerClass(lr=5e-5, lr_dropout=lr_dropout, clipnorm=clipnorm, name='src_dst_opt') self.src_dst_opt = OptimizerClass(lr=lr, lr_dropout=lr_dropout, clipnorm=clipnorm, name='src_dst_opt')
self.src_dst_opt.initialize_variables (self.G_weights, vars_on_cpu=optimizer_vars_on_cpu) self.src_dst_opt.initialize_variables (self.G_weights, vars_on_cpu=optimizer_vars_on_cpu)
self.model_filename_list += [ (self.src_dst_opt, 'src_dst_opt.npy') ] self.model_filename_list += [ (self.src_dst_opt, 'src_dst_opt.npy') ]
if gan_power != 0: if gan_power != 0:
if self.options['gan_version'] == 2: if self.options['gan_version'] == 2:
self.GAN_opt = OptimizerClass(lr=5e-5, lr_dropout=lr_dropout, lr_cos=lr_cos, clipnorm=clipnorm, name='D_src_dst_opt') self.GAN_opt = OptimizerClass(lr=lr, lr_dropout=lr_dropout, lr_cos=lr_cos, clipnorm=clipnorm, name='D_src_dst_opt')
self.GAN_opt.initialize_variables ( self.GAN.get_weights(), vars_on_cpu=optimizer_vars_on_cpu, lr_dropout_on_cpu=self.options['lr_dropout']=='cpu')#+self.D_src_x2.get_weights() self.GAN_opt.initialize_variables ( self.GAN.get_weights(), vars_on_cpu=optimizer_vars_on_cpu, lr_dropout_on_cpu=self.options['lr_dropout']=='cpu')#+self.D_src_x2.get_weights()
self.model_filename_list += [ (self.GAN_opt, 'D_src_v2_opt.npy') ] self.model_filename_list += [ (self.GAN_opt, 'D_src_v2_opt.npy') ]
else: else:
self.GAN_opt = OptimizerClass(lr=5e-5, lr_dropout=lr_dropout, lr_cos=lr_cos, clipnorm=clipnorm, name='GAN_opt') self.GAN_opt = OptimizerClass(lr=lr, lr_dropout=lr_dropout, lr_cos=lr_cos, clipnorm=clipnorm, name='GAN_opt')
self.GAN_opt.initialize_variables ( self.GAN.get_weights(), vars_on_cpu=optimizer_vars_on_cpu, lr_dropout_on_cpu=self.options['lr_dropout']=='cpu')#+self.D_src_x2.get_weights() self.GAN_opt.initialize_variables ( self.GAN.get_weights(), vars_on_cpu=optimizer_vars_on_cpu, lr_dropout_on_cpu=self.options['lr_dropout']=='cpu')#+self.D_src_x2.get_weights()
self.model_filename_list += [ (self.GAN_opt, 'GAN_opt.npy') ] self.model_filename_list += [ (self.GAN_opt, 'GAN_opt.npy') ]
@ -741,7 +753,7 @@ class AMPModel(ModelBase):
random_ct_samples_path=training_data_dst_path if ct_mode is not None else None #and not self.pretrain random_ct_samples_path=training_data_dst_path if ct_mode is not None else None #and not self.pretrain
cpu_count = multiprocessing.cpu_count() cpu_count = min(multiprocessing.cpu_count(), self.options['cpu_cap'])
src_generators_count = cpu_count // 2 src_generators_count = cpu_count // 2
dst_generators_count = cpu_count // 2 dst_generators_count = cpu_count // 2
if ct_mode is not None: if ct_mode is not None:
@ -762,7 +774,6 @@ class AMPModel(ModelBase):
'random_blur': self.options['random_blur'], 'random_blur': self.options['random_blur'],
'random_jpeg': self.options['random_jpeg'], 'random_jpeg': self.options['random_jpeg'],
'transform':True, 'channel_type' : channel_type, 'ct_mode': ct_mode, 'transform':True, 'channel_type' : channel_type, 'ct_mode': ct_mode,
'random_hsv_shift_amount' : random_hsv_power,
'face_type':self.face_type, 'data_format':nn.data_format, 'resolution': resolution}, 'face_type':self.face_type, 'data_format':nn.data_format, 'resolution': resolution},
{'sample_type': SampleProcessor.SampleType.FACE_IMAGE,'warp':False, {'sample_type': SampleProcessor.SampleType.FACE_IMAGE,'warp':False,
'transform':True, 'channel_type' : channel_type, 'ct_mode': ct_mode, 'transform':True, 'channel_type' : channel_type, 'ct_mode': ct_mode,
@ -916,28 +927,40 @@ class AMPModel(ModelBase):
target_srcm, target_dstm = [ nn.to_data_format(x,"NHWC", self.model_data_format) for x in ([target_srcm, target_dstm] )] target_srcm, target_dstm = [ nn.to_data_format(x,"NHWC", self.model_data_format) for x in ([target_srcm, target_dstm] )]
n_samples = min(4, self.get_batch_size(), 800 // self.resolution ) n_samples = min(self.get_batch_size(), self.options['preview_samples'])
result = [] result = []
i = np.random.randint(n_samples) if not for_history else 0 #i = np.random.randint(n_samples) if not for_history else 0
if filenames is not None and len(filenames) > 0: for i in range(n_samples if not for_history else 1):
S[i] = label_face_filename(S[i], filenames[0][i]) if filenames is not None and len(filenames) > 0:
D[i] = label_face_filename(D[i], filenames[1][i]) S[i] = label_face_filename(S[i], filenames[0][i])
D[i] = label_face_filename(D[i], filenames[1][i])
st = [ np.concatenate ((S[i], D[i], DD[i]*DDM_000[i]), axis=1) ] st = []
st += [ np.concatenate ((SS[i], DD[i], SD_100[i] ), axis=1) ] temp_r = []
for i in range(n_samples if not for_history else 1):
result += [ ('AMP morph 1.0', np.concatenate (st, axis=0 )), ] st = [ np.concatenate ((S[i], SS[i], D[i]), axis=1) ]
st += [ np.concatenate ((DD[i], DD[i]*DDM_000[i], SD_100[i] ), axis=1) ]
st = [ np.concatenate ((DD[i], SD_025[i], SD_050[i]), axis=1) ] temp_r += [ np.concatenate (st, axis=1) ]
st += [ np.concatenate ((SD_065[i], SD_075[i], SD_100[i]), axis=1) ] result += [ ('AMP morph 1.0', np.concatenate (temp_r, axis=0 )), ]
result += [ ('AMP morph list', np.concatenate (st, axis=0 )), ] # result += [ ('AMP morph 1.0', np.concatenate (st, axis=0 )), ]
st = []
st = [ np.concatenate ((DD[i], SD_025[i]*DDM_025[i]*SDM_025[i], SD_050[i]*DDM_050[i]*SDM_050[i]), axis=1) ] temp_r = []
st += [ np.concatenate ((SD_065[i]*DDM_065[i]*SDM_065[i], SD_075[i]*DDM_075[i]*SDM_075[i], SD_100[i]*DDM_100[i]*SDM_100[i]), axis=1) ] for i in range(n_samples if not for_history else 1):
result += [ ('AMP morph list masked', np.concatenate (st, axis=0 )), ] st = [ np.concatenate ((DD[i], SD_025[i], SD_050[i]), axis=1) ]
st += [ np.concatenate ((SD_065[i], SD_075[i], SD_100[i]), axis=1) ]
temp_r += [ np.concatenate (st, axis=1) ]
result += [ ('AMP morph list', np.concatenate (temp_r, axis=0 )), ]
#result += [ ('AMP morph list', np.concatenate (st, axis=0 )), ]
st = []
temp_r = []
for i in range(n_samples if not for_history else 1):
st = [ np.concatenate ((DD[i], SD_025[i]*DDM_025[i]*SDM_025[i], SD_050[i]*DDM_050[i]*SDM_050[i]), axis=1) ]
st += [ np.concatenate ((SD_065[i]*DDM_065[i]*SDM_065[i], SD_075[i]*DDM_075[i]*SDM_075[i], SD_100[i]*DDM_100[i]*SDM_100[i]), axis=1) ]
temp_r += [ np.concatenate (st, axis=1) ]
result += [ ('AMP morph list masked', np.concatenate (temp_r, axis=0 )), ]
#result += [ ('AMP morph list masked', np.concatenate (st, axis=0 )), ]
return result return result

26
models/Model_AMP/config_schema.json
View file

@ -4,11 +4,16 @@
"definitions": { "definitions": {
"dfl_config": { "dfl_config": {
"type": "object", "type": "object",
"additionalProperties": false, "additionalProperties": true,
"properties": { "properties": {
"use_fp16": { "use_fp16": {
"type": "boolean" "type": "boolean"
}, },
"cpu_cap": {
"type": "integer",
"minimum": 1,
"maximum": 256
},
"morph_factor": { "morph_factor": {
"type": "number", "type": "number",
"minimum":0.0, "minimum":0.0,
@ -97,14 +102,14 @@
"MS-SSIM+L1" "MS-SSIM+L1"
] ]
}, },
"lr_modifier": {
"type": "integer",
"minimum": -100,
"maximum": 100
},
"random_warp": { "random_warp": {
"type": "boolean" "type": "boolean"
}, },
"random_hsv_power": {
"type": "number",
"minimum": 0.0,
"maximum": 0.3
},
"random_downsample": { "random_downsample": {
"type": "boolean" "type": "boolean"
}, },
@ -169,6 +174,11 @@
"random_dst_flip": { "random_dst_flip": {
"type": "boolean" "type": "boolean"
}, },
"preview_samples": {
"type": "integer",
"minimum": 1,
"maximum": 64
},
"batch_size": { "batch_size": {
"type": "integer", "type": "integer",
"minimum": 1 "minimum": 1
@ -205,6 +215,7 @@
} }
}, },
"required": [ "required": [
"cpu_cap",
"adabelief", "adabelief",
"ae_dims", "ae_dims",
"autobackup_hour", "autobackup_hour",
@ -228,16 +239,15 @@
"gan_version", "gan_version",
"loss_function", "loss_function",
"lr_dropout", "lr_dropout",
"lr_modifier",
"masked_training", "masked_training",
"maximum_n_backups", "maximum_n_backups",
"models_opt_on_gpu", "models_opt_on_gpu",
"mouth_prio", "mouth_prio",
"pretrain",
"random_blur", "random_blur",
"random_color", "random_color",
"random_downsample", "random_downsample",
"random_dst_flip", "random_dst_flip",
"random_hsv_power",
"random_jpeg", "random_jpeg",
"random_noise", "random_noise",
"random_src_flip", "random_src_flip",

32
models/Model_SAEHD/Model.py
View file

@ -58,7 +58,6 @@ class SAEHDModel(ModelBase):
default_loss_function = self.options['loss_function'] = self.load_or_def_option('loss_function', 'SSIM') default_loss_function = self.options['loss_function'] = self.load_or_def_option('loss_function', 'SSIM')
default_random_warp = self.options['random_warp'] = self.load_or_def_option('random_warp', True) default_random_warp = self.options['random_warp'] = self.load_or_def_option('random_warp', True)
default_random_hsv_power = self.options['random_hsv_power'] = self.load_or_def_option('random_hsv_power', 0.0)
default_random_downsample = self.options['random_downsample'] = self.load_or_def_option('random_downsample', False) default_random_downsample = self.options['random_downsample'] = self.load_or_def_option('random_downsample', False)
default_random_noise = self.options['random_noise'] = self.load_or_def_option('random_noise', False) default_random_noise = self.options['random_noise'] = self.load_or_def_option('random_noise', False)
default_random_blur = self.options['random_blur'] = self.load_or_def_option('random_blur', False) default_random_blur = self.options['random_blur'] = self.load_or_def_option('random_blur', False)
@ -72,7 +71,10 @@ class SAEHDModel(ModelBase):
default_random_color = self.options['random_color'] = self.load_or_def_option('random_color', False) default_random_color = self.options['random_color'] = self.load_or_def_option('random_color', False)
default_clipgrad = self.options['clipgrad'] = self.load_or_def_option('clipgrad', False) default_clipgrad = self.options['clipgrad'] = self.load_or_def_option('clipgrad', False)
default_pretrain = self.options['pretrain'] = self.load_or_def_option('pretrain', False) default_pretrain = self.options['pretrain'] = self.load_or_def_option('pretrain', False)
#default_use_fp16 = self.options['use_fp16'] = self.load_or_def_option('use_fp16', False) default_cpu_cap = self.options['cpu_cap'] = self.load_or_def_option('cpu_cap', 8)
default_preview_samples = self.options['preview_samples'] = self.load_or_def_option('preview_samples', 4)
default_full_preview = self.options['force_full_preview'] = self.load_or_def_option('force_full_preview', False)
default_lr_modifier = self.options['lr_modifier'] = self.load_or_def_option('lr_modifier', 0)
ask_override = False if self.read_from_conf else self.ask_override() ask_override = False if self.read_from_conf else self.ask_override()
if self.is_first_run() or ask_override: if self.is_first_run() or ask_override:
@ -81,12 +83,15 @@ class SAEHDModel(ModelBase):
self.ask_autobackup_hour() self.ask_autobackup_hour()
self.ask_maximum_n_backups() self.ask_maximum_n_backups()
self.ask_write_preview_history() self.ask_write_preview_history()
self.options['preview_samples'] = np.clip ( io.input_int ("Number of samples to preview", default_preview_samples, add_info="1 - 16", help_message="Typical fine value is 4"), 1, 16 )
self.ask_target_iter() self.ask_target_iter()
self.ask_retraining_samples() self.ask_retraining_samples()
self.ask_random_src_flip() self.ask_random_src_flip()
self.ask_random_dst_flip() self.ask_random_dst_flip()
self.ask_batch_size(suggest_batch_size) self.ask_batch_size(suggest_batch_size)
self.options['use_fp16'] = io.input_bool ("Use fp16", default_usefp16, help_message='Increases training/inference speed, reduces model size. Model may crash. Enable it after 1-5k iters.') self.options['use_fp16'] = io.input_bool ("Use fp16", default_usefp16, help_message='Increases training/inference speed, reduces model size. Model may crash. Enable it after 1-5k iters.')
self.options['cpu_cap'] = np.clip ( io.input_int ("Max cpu cores to use.", default_cpu_cap, add_info="1 - 256", help_message="Typical fine value is 8"), 1, 256 )
if self.is_first_run(): if self.is_first_run():
if (self.read_from_conf and not self.config_file_exists) or not self.read_from_conf: if (self.read_from_conf and not self.config_file_exists) or not self.read_from_conf:
@ -176,11 +181,11 @@ class SAEHDModel(ModelBase):
self.options['loss_function'] = io.input_str(f"Loss function", default_loss_function, ['SSIM', 'MS-SSIM', 'MS-SSIM+L1'], self.options['loss_function'] = io.input_str(f"Loss function", default_loss_function, ['SSIM', 'MS-SSIM', 'MS-SSIM+L1'],
help_message="Change loss function used for image quality assessment.") help_message="Change loss function used for image quality assessment.")
self.options['lr_modifier'] = np.clip (io.input_int("Learningrate factor", default_lr_modifier, add_info="-100 .. 100", help_message="Modify the Learning rate: 100 == multipy by 4, -100 == divide by 4"), -100, 100)
self.options['random_warp'] = io.input_bool ("Enable random warp of samples", default_random_warp, help_message="Random warp is required to generalize facial expressions of both faces. When the face is trained enough, you can disable it to get extra sharpness and reduce subpixel shake for less amount of iterations.") self.options['random_warp'] = io.input_bool ("Enable random warp of samples", default_random_warp, help_message="Random warp is required to generalize facial expressions of both faces. When the face is trained enough, you can disable it to get extra sharpness and reduce subpixel shake for less amount of iterations.")
self.options['random_hsv_power'] = np.clip ( io.input_number ("Random hue/saturation/light intensity", default_random_hsv_power, add_info="0.0 .. 0.3", help_message="Random hue/saturation/light intensity applied to the src face set only at the input of the neural network. Stabilizes color perturbations during face swapping. Reduces the quality of the color transfer by selecting the closest one in the src faceset. Thus the src faceset must be diverse enough. Typical fine value is 0.05"), 0.0, 0.3 )
self.options['random_downsample'] = io.input_bool("Enable random downsample of samples", default_random_downsample, help_message="") self.options['random_downsample'] = io.input_bool("Enable random downsample of samples", default_random_downsample, help_message="")
self.options['random_noise'] = io.input_bool("Enable random noise added to samples", default_random_noise, help_message="") self.options['random_noise'] = io.input_bool("Enable random noise added to samples", default_random_noise, help_message="")
self.options['random_blur'] = io.input_bool("Enable random blur of samples", default_random_blur, help_message="") self.options['random_blur'] = io.input_bool("Enable random blur of samples", default_random_blur, help_message="")
@ -270,14 +275,12 @@ class SAEHDModel(ModelBase):
random_warp = False if self.pretrain else self.options['random_warp'] random_warp = False if self.pretrain else self.options['random_warp']
random_src_flip = self.random_src_flip if not self.pretrain else True random_src_flip = self.random_src_flip if not self.pretrain else True
random_dst_flip = self.random_dst_flip if not self.pretrain else True random_dst_flip = self.random_dst_flip if not self.pretrain else True
random_hsv_power = self.options['random_hsv_power'] if not self.pretrain else 0.0
blur_out_mask = self.options['blur_out_mask'] blur_out_mask = self.options['blur_out_mask']
if self.pretrain: if self.pretrain:
self.options_show_override['lr_dropout'] = 'n' self.options_show_override['lr_dropout'] = 'n'
self.options_show_override['random_warp'] = False self.options_show_override['random_warp'] = False
self.options_show_override['gan_power'] = 0.0 self.options_show_override['gan_power'] = 0.0
self.options_show_override['random_hsv_power'] = 0.0
self.options_show_override['face_style_power'] = 0.0 self.options_show_override['face_style_power'] = 0.0
self.options_show_override['bg_style_power'] = 0.0 self.options_show_override['bg_style_power'] = 0.0
self.options_show_override['uniform_yaw'] = True self.options_show_override['uniform_yaw'] = True
@ -360,7 +363,14 @@ class SAEHDModel(ModelBase):
self.model_filename_list += [ [self.D_src, 'GAN.npy'] ] self.model_filename_list += [ [self.D_src, 'GAN.npy'] ]
# Initialize optimizers # Initialize optimizers
lr=5e-5 lr_modifier = self.options['lr_modifier']
if lr_modifier == 0:
lr = 5e-5
elif lr_modifier < 0:
lr = 5e-5 / abs( lr_modifier * 4/100 )
else:
lr = 5e-5 * abs( lr_modifier * 4/100 )
if self.options['lr_dropout'] in ['y','cpu'] and not self.pretrain: if self.options['lr_dropout'] in ['y','cpu'] and not self.pretrain:
lr_cos = 500 lr_cos = 500
lr_dropout = 0.3 lr_dropout = 0.3
@ -795,7 +805,7 @@ class SAEHDModel(ModelBase):
random_ct_samples_path=training_data_dst_path if ct_mode is not None and not self.pretrain else None random_ct_samples_path=training_data_dst_path if ct_mode is not None and not self.pretrain else None
cpu_count = multiprocessing.cpu_count() cpu_count = min(multiprocessing.cpu_count(), self.options['cpu_cap'])
src_generators_count = cpu_count // 2 src_generators_count = cpu_count // 2
dst_generators_count = cpu_count // 2 dst_generators_count = cpu_count // 2
if ct_mode is not None: if ct_mode is not None:
@ -815,7 +825,7 @@ class SAEHDModel(ModelBase):
'random_noise': self.options['random_noise'], 'random_noise': self.options['random_noise'],
'random_blur': self.options['random_blur'], 'random_blur': self.options['random_blur'],
'random_jpeg': self.options['random_jpeg'], 'random_jpeg': self.options['random_jpeg'],
'transform':True, 'channel_type' : channel_type, 'ct_mode': ct_mode, 'random_hsv_shift_amount' : random_hsv_power, 'transform':True, 'channel_type' : channel_type, 'ct_mode': ct_mode,
'face_type':self.face_type, 'data_format':nn.data_format, 'resolution': resolution}, 'face_type':self.face_type, 'data_format':nn.data_format, 'resolution': resolution},
{'sample_type': SampleProcessor.SampleType.FACE_IMAGE,'warp':False , 'transform':True, 'channel_type' : channel_type, 'ct_mode': ct_mode, 'face_type':self.face_type, 'data_format':nn.data_format, 'resolution': resolution}, {'sample_type': SampleProcessor.SampleType.FACE_IMAGE,'warp':False , 'transform':True, 'channel_type' : channel_type, 'ct_mode': ct_mode, 'face_type':self.face_type, 'data_format':nn.data_format, 'resolution': resolution},
{'sample_type': SampleProcessor.SampleType.FACE_MASK, 'warp':False , 'transform':True, 'channel_type' : SampleProcessor.ChannelType.G, 'face_mask_type' : SampleProcessor.FaceMaskType.FULL_FACE, 'face_type':self.face_type, 'data_format':nn.data_format, 'resolution': resolution}, {'sample_type': SampleProcessor.SampleType.FACE_MASK, 'warp':False , 'transform':True, 'channel_type' : SampleProcessor.ChannelType.G, 'face_mask_type' : SampleProcessor.FaceMaskType.FULL_FACE, 'face_type':self.face_type, 'data_format':nn.data_format, 'resolution': resolution},
@ -965,14 +975,14 @@ class SAEHDModel(ModelBase):
target_srcm, target_dstm = [ nn.to_data_format(x,"NHWC", self.model_data_format) for x in ([target_srcm, target_dstm] )] target_srcm, target_dstm = [ nn.to_data_format(x,"NHWC", self.model_data_format) for x in ([target_srcm, target_dstm] )]
n_samples = min(4, self.get_batch_size(), 800 // self.resolution ) n_samples = min(self.get_batch_size(), self.options['preview_samples'])
if filenames is not None and len(filenames) > 0: if filenames is not None and len(filenames) > 0:
for i in range(n_samples): for i in range(n_samples):
S[i] = label_face_filename(S[i], filenames[0][i]) S[i] = label_face_filename(S[i], filenames[0][i])
D[i] = label_face_filename(D[i], filenames[1][i]) D[i] = label_face_filename(D[i], filenames[1][i])
if self.resolution <= 256: if self.resolution <= 256 or self.options['force_full_preview'] == True:
result = [] result = []
st = [] st = []

29
models/Model_SAEHD/config_schema.json
View file

@ -4,7 +4,7 @@
"definitions": { "definitions": {
"dfl_config": { "dfl_config": {
"type": "object", "type": "object",
"additionalProperties": false, "additionalProperties": true,
"properties": { "properties": {
"use_fp16": { "use_fp16": {
"type": "boolean" "type": "boolean"
@ -13,6 +13,11 @@
"type": "string", "type": "string",
"pattern": "^(df|liae)-(\\b(?!\\w*(\\w)\\w*\\1)[udtc]+\\b)+|^(df|liae)$" "pattern": "^(df|liae)-(\\b(?!\\w*(\\w)\\w*\\1)[udtc]+\\b)+|^(df|liae)$"
}, },
"cpu_cap": {
"type": "integer",
"minimum": 1,
"maximum": 256
},
"resolution": { "resolution": {
"type": "integer", "type": "integer",
"minimum": 64, "minimum": 64,
@ -90,14 +95,14 @@
"MS-SSIM+L1" "MS-SSIM+L1"
] ]
}, },
"lr_modifier": {
"type": "integer",
"minimum": -100,
"maximum": 100
},
"random_warp": { "random_warp": {
"type": "boolean" "type": "boolean"
}, },
"random_hsv_power": {
"type": "number",
"minimum": 0.0,
"maximum": 0.3
},
"random_downsample": { "random_downsample": {
"type": "boolean" "type": "boolean"
}, },
@ -150,6 +155,14 @@
"pretrain": { "pretrain": {
"type": "boolean" "type": "boolean"
}, },
"preview_samples": {
"type": "integer",
"minimum": 1,
"maximum": 64
},
"force_full_preview": {
"type": "boolean"
},
"session_name": { "session_name": {
"type": "string" "type": "string"
}, },
@ -213,6 +226,7 @@
} }
}, },
"required": [ "required": [
"cpu_cap",
"adabelief", "adabelief",
"ae_dims", "ae_dims",
"archi", "archi",
@ -237,16 +251,17 @@
"gan_version", "gan_version",
"loss_function", "loss_function",
"lr_dropout", "lr_dropout",
"lr_modifier",
"masked_training", "masked_training",
"maximum_n_backups", "maximum_n_backups",
"models_opt_on_gpu", "models_opt_on_gpu",
"mouth_prio", "mouth_prio",
"pretrain", "pretrain",
"preview_samples",
"random_blur", "random_blur",
"random_color", "random_color",
"random_downsample", "random_downsample",
"random_dst_flip", "random_dst_flip",
"random_hsv_power",
"random_jpeg", "random_jpeg",
"random_noise", "random_noise",
"random_src_flip", "random_src_flip",

2
models/Model_XSeg/config_schema.json
View file

@ -31,7 +31,7 @@
"required": [ "required": [
"batch_size", "batch_size",
"face_type", "face_type",
"pretrain", "pretrain"
], ],
"title": "dfl_config" "title": "dfl_config"
} }

2
samplelib/SampleGeneratorFaceXSeg.py
View file

@ -209,7 +209,7 @@ class SegmentedSampleFilterSubprocessor(Subprocessor):
#override #override
def process_info_generator(self): def process_info_generator(self):
for i in range(multiprocessing.cpu_count()): for i in range(min(multiprocessing.cpu_count(),8)):
yield 'CPU%d' % (i), {}, {'samples':self.samples, 'count_xseg_mask':self.count_xseg_mask} yield 'CPU%d' % (i), {}, {'samples':self.samples, 'count_xseg_mask':self.count_xseg_mask}
#override #override